Google Bets Cheaper Gemini AI Models Can Close Enterprise Gap as Alphabet Wins Antitrust Relief

Alphabet is leaning on cheaper, faster artificial intelligence models, a massive Google Cloud customer base, and the continued strength of its advertising business as it tries to narrow the gap with AI leaders OpenAI and Anthropic in the enterprise market.

The company is positioning Gemini 3.8 Flash as a key part of that strategy, with the latest model focused heavily on coding, reasoning and agentic tasks that Google sees as an important part of its businesses seeking to turn AI advances into measurable productivity gains.

Google DeepMind describes Gemini 3.8 Flash as its strongest Flash model yet for reasoning and coding, with significant improvements over Gemini 3.7 Flash in software engineering and multi-step tasks.

Tulsee Doshi, senior director of product management at Google DeepMind, told CNBC that the recent Flash models had “really surprised us in positive ways in their performance,” creating opportunities for Google to expand their use.

The economics are central to Google’s pitch.

Gemini 3.8 Flash is priced at 75 cents per million input tokens and $3.75 per million output tokens, matching the introductory price of its previous Flash model even as Google claims substantial gains in coding, reasoning and agentic capabilities.

Smaller models are cheaper to operate and can be deployed and improved more rapidly than Google’s largest frontier systems. Their increasing ability to handle complex tasks also gives Google a way to offer businesses AI capabilities without forcing them to pay for the most expensive models. It is expected to become more relevant as companies move from experimenting with generative AI to deploying autonomous agents that perform multistep tasks and consume AI inference at much larger volumes.

Still, Google faces a significant enterprise-market challenge.

“From a product perspective this model seems to keep Google in the race, but probably won’t change the fact they are a distant third in the enterprise market,” said Gil Luria, an analyst at D.A. Davidson, who recommends holding Alphabet shares.

Google Targets Microsoft and Anthropic on Price

Google is also trying to compete by changing how businesses pay for AI. Gemini Enterprise is adding pay-as-you-go pricing, token discounts of as much as 20%, monthly limits on agent spending and a zero-dollar base subscription option. Google has also sought to distinguish its offering from Microsoft and Anthropic by arguing that recurring seat fees and separate product licenses can make competing AI products more expensive and less flexible.

The approach relies partly on Google’s existing distribution.

Nearly three-quarters of Google Cloud customers are already using the company’s AI products, according to Google. Google Cloud CEO Thomas Kurian told CNBC that those customers are spending roughly 50% more than their original commitments.

Analysts predict that the installed base could give Alphabet an advantage that does not depend entirely on having the industry’s single best AI model. Google can package Gemini into Cloud, Workspace, Android, Search and other products that businesses already use, potentially lowering the cost of adopting AI.

Demis Hassabis, the head of Google DeepMind, outlined an even broader vision at the G20 Innovation meeting on Wednesday, saying Gemini could operate as a general-purpose layer that coordinates cheaper, specialized models and AI agents.

Many expect that approach to shift the competitive equation. Rather than relying exclusively on one frontier model outperforming every rival, Google could use its breadth of products, computing infrastructure and distribution to coordinate a network of models and agents.

Hassabis spoke publicly for the first time since DeepMind’s reorganization last month, under which he moved from chief executive to chairman of the unit.

Cybersecurity Becomes Another AI Opportunity

Google is extending the same cost argument into cybersecurity with Gemini 3.8 Flash Cyber.

The company says the model can identify and patch software vulnerabilities at frontier-level performance while operating faster and at lower cost than larger AI systems.

“We’re really excited about being able to provide an offering to defenders that is a fraction of the cost, much faster, while still showcasing that frontier-level performance,” Doshi said.

Because the same capabilities could potentially be exploited by attackers, Google is initially restricting access to a small group of trusted government and enterprise cybersecurity defenders through its Fairwind Program.

Massive AI Spending Raises the Stakes

Google is spending heavily on data centers, computing capacity and AI development. That creates a financial imperative for Gemini to gain market share and generate higher cloud consumption, enterprise software revenue and advertising opportunities.

Warren Buffett’s successor at Berkshire Hathaway, Greg Abel, said Wednesday that the conglomerate views Alphabet as an AI winner, partly because of what its portfolio companies are seeing from Google’s technology.

“We have a lot of visibility from within our companies as to how we’re using AI, what type of benefits it’s delivering, so that brought incremental interest, and then we saw Google as a significant player,” Abel told CNBC’s Becky Quick.

Alphabet also continues to have a powerful source of cash to finance that AI push: advertising.

Google’s advertising business grew 14% in the latest quarter, providing the company with a large and relatively mature cash-generating operation that can help fund its expansion into AI.

Antitrust Rulings Remove a Major Constraint

Alphabet received another boost Wednesday when a federal judge overseeing the U.S. Justice Department’s ad-tech antitrust case ruled that Google would not have to sell its AdX advertising exchange. The court instead opted for behavioral remedies rather than the structural breakup sought by the government.

The decision follows a separate antitrust ruling last year in which a judge rejected demands that Google divest its Chrome browser.

The two rulings do not eliminate Google’s regulatory risks, but they reduce the immediate threat of a forced restructuring of some of its most important businesses.

Antitrust attorney Wyatt Fore, a partner at Shinder Cantor Lerner, described the latest decision as a “big deal,” noting that Google is entering the AI race without the structural restrictions that a breakup could have imposed.

For Alphabet, that matters strategically. The company can continue combining its AI models with Search, Cloud, advertising, Workspace and Android while deploying its infrastructure at scale. The central investment debate is therefore shifting from whether Google can remain relevant in AI to whether it can convert that enormous ecosystem advantage into durable enterprise market share.